Looking for real eyes to examine

[jfutral]

Wow. It has been almost 10 years since I was last around here. Nice to see that this place is still vibrant. Gotta love the DIY generation!

I'm looking for feedback on this amp/schematic (as much as one can give feedback from a drawing) I just built for a friend. I'm no designer. I might be able to pass as a Suzuki violin version of an amp builder. You know, the kind who can read the notes, but has no idea what he is actually playing. I know math, I just don't KNOW this math like I probably should.

Since this is going out for someone other than me and he is very far away, I want to make sure I am not sending something out that will eventually cause a fire and hopefully will last him until he is rich and famous and can buy a real amp from any number of great boutique craftsmen out there.

I made this amp before as a head with 2-6L6s and loved it, so I thought I'd try it with 2-6v6s. It's a bit of a frankenstein. But, I have to say I very much like the way this amp sounds. I may have to build one for myself.

Essentially, it is a hybrid single input, switchable (not indicated on the drawing) preamp with A as a typical Fender pre and B from a typical Marshall pre. The power is pseudo Plexi, except I went with a Cathode bias design rather than fixed. (My PT does not have a bias tap). I had an old chassis with some things still in place from an old build, I used what parts I had around and picked up a few extra to fill in the holes.

I wrote in the voltages on the drawing everywhere I thought they might matter. My big concern is the plate dissipation of the output tubes. I hear cathode bias supposedly can run hotter, but this hot? If not, how would I get the current down without pumping up the cathode resistor anymore? Just seems large even now. But, like I said, I am not a designer, so I don't know. Nothing is glowing. So I guess I'm OK?

Also, at first I had a little issue with the "Fendish" side being louder than the "Marshish" side, especially at low volumes. I made several feeble attempts, but eventually took out the 100k mix resistor before the PI on the Marshish side. That seemed to even out the volumes without any noticeable negative affect. Or maybe something else happened at the same time that I am unaware of. Regardless, the balance works now. Maybe there is a better way.

The only other thing I am otherwise dealing with is a low level hum that doesn't really make itself noticeable until the MV is above 7. The Fendish side can make it a bit louder if it is cranked above 7 at the same time. He probably will never drive it that loud, but I want to make sure this isn't symptomatic of something else. The Marshish side is fairly quiet, depending on what preamp tubes I throw in there.

I appreciate any feedback. Let me know if you need any other information. She is laid bare before me.

Thanks,
Joe

[txstrat]

The 320ohms cathode resistor on the power tubes is fine IMHO. Fender even used 250 ohms with two 6V6s and the amps still live. Cathode biased amps get hotter because the tubes are conducting the full 360 degrees. That's why I usually put them in stand by, when I don't play (I use 270 ohms).
I once forgot this and after a couple hours the amp got so hot you couldn't touch the chassis (it even smelled funny). I'd also put the 100k mixing resistor back in, because it might lead to issues when the treble pot (Marshish channel) is fully up or down. To get that channel louder you could decrease the cathode resistor on the first triode to 1.5K or so. But that would lead you to more gain also. IMHO there's no need to get the channels equal in volume, since you have independent volume controls.
Hum is often caused by bad grounding. Do you use a star ground or something similar?

[Robert M. Martinelli]

Just as a footnote to what Matt said...

According to your figures, you have 387.5-25.5=362V dropping across each 6V6 in quiescent conditions, the total quiescent current flowing through the cathode bias resistor is 25.5/320=about 79.7mA, this means (theoretically, in practice the two values will somewhat differ) that each tube contributes with 39.85 mA, of which 1V/470Ohm=2.12 mA screen current and the rest (37.7mA) plate current. The quiescent power dissipation for each 6V6 is then 362*0.0377=13.64W. 6V6s are rated at 14W, so it seems your amp is biased very close to 100%. Luckily, the cathode bias arrangement is somewhat "self-regulating" (as the current across the tubes increases, so does the voltage drop across the cathode resistor, elevating the cathode voltage, thus shifting the bias to the colder side and giving the tubes some relief (and introducing some "cathode bias sag" too). It's true that some amps are even biased hotter than that (e.g. AC30,which is very rough on its output tubes) but, especially with modern tubes, I would bias the thing a little bit colder (70-80% of the max Pd, which should be enough to keep the cross-over distortion at bay with no sonic drawbacks IMHO).

A last word about the cathode bias resistor. According to your figures, it has to continuously dissipate a minimum of 25.5V*0.0797mA=Slightly more than 2W, so, to keep it at a reasonable operating temperature, I'd use at least a 5W (or better yet a 7W) one.

Hope this helps

Best regards

Bob

[tubeswell]

I wrote in the voltages on the drawing everywhere I thought they might matter. My big concern is the plate dissipation of the output tubes. I hear cathode bias supposedly can run hotter, but this hot?

each tube is dissipating about 14W (including the screens). I wouldn't run it any harder, but with good rugged 6V6GT (or GTA) or JJ6V6S, methinks it'll be fine

If not, how would I get the current down without pumping up the cathode resistor anymore? Just seems large even now. But, like I said, I am not a designer, so I don't know. Nothing is glowing. So I guess I'm OK?

Dropping the B+ is another way (try zenering down the HT winding with some reverse-biased zeners between the CT and the ground - 2 x 15V 5W zeners in series ought to do it - banded ends pointing to ground. Then you could most likely change the cathode resistor back to 250-270R)

Also, at first I had a little issue with the "Fendish" side being louder than the "Marshish" side, especially at low volumes. I made several feeble attempts, but eventually took out the 100k mix resistor before the PI on the Marshish side. That seemed to even out the volumes without any noticeable negative affect. Or maybe something else happened at the same time that I am unaware of. Regardless, the balance works now. Maybe there is a better way.

Your 'marshall' side has a big lossy tone stack after the CF stage, compared to your BF fender side, which has the recovery stage after the tone stack. one way woudl be to add another stage after the marshall tone stack - you would seem to have a spare triode

The only other thing I am otherwise dealing with is a low level hum that doesn't really make itself noticeable until the MV is above 7. The Fendish side can make it a bit louder if it is cranked above 7 at the same time. He probably will never drive it that loud, but I want to make sure this isn't symptomatic of something else. The Marshish side is fairly quiet, depending on what preamp tubes I throw in there

How are the ground returns sorted out in the pre-amp? (Have you read Merlin's article here? The Valve Wizard )

[Ronsonic]

You will want to have a coupling cap (DC blocking cap) after each tone stack you'll see the wiper of the Marshish treble pot is directly coupled to the plate of the fendish side. Gonna be a little noisy. I'd give both sides a mix resistor too.

[jfutral]

I had forgotten all about star grounding until just the other day. So, no. No star ground. I am using a turret board and created a sort of bus along one side of it, grounded on the end furthest from the filter caps and PT.

Joe

[jfutral]

I have not read his article. Didn't even know it was out there. Great read. I did find one ground loop/redundant ground. I wish eliminating it had helped. Almost all the grounds for the preamp are along a bus I wired along one side of the turret board. Then it ties to ground at the end furthest from the filter caps and PT. The only preamp grounds that aren't along that line are the cathode C/Rs which I have at the tubes. The heaters are all wired for 12.6v so no ground on pin 9. I do have 100k resisters to ground at the PT for the heaters.

I'll have to re-read the article. I read it this morning and have been examining, re-wiring and grounding, re-examining all day. My eyes hurt! I think I made things worse. Grrr. Maybe I should have been happy with what I had. The hum wasn't really noticeable until pushing the MV and channel volumes past 7. Sometimes 70% is good enough! :-) But the lure of perfection seems so attainable! Well, I did at least find a few weak solders and a couple weak wiring joints. Probably would have given out right when my friend was to jam his "Eric Johnson" style solo. That would have been embarrassing.

Saving that last triode for when my friend says "Hey. Can you make it do this?". But he may just ask "Hey. Can you make that channel louder?" My luck.

Thanks,
Joe

[jfutral]

Coupling cap! Yes! I did leave that out. Good catch. At it did help some of the noise.

Thanks,
Joe

[tubeswell]

Hi Joe

I have found that doing grounding that way results in extremely quiet amps, but for it to work best, you have to have all ground returns, each running by their own separate wire, to the same ground return wire as the relevant filter cap ground that is associated with that part of the circuit. Ground busses defeat the purpose. The attached schematic is for a PR clone I finished recently that was wired in this way. (Note that I put in separate filter caps and did away with the cap can). The only aspect that is not clear on this schematic is that I kept the pre-amp ground returns (except for the RT and driver) split from the output stage ground returns. That is; the 'pre-amp' filter cap grounds and their associated circuit ground returns are routed to the input socket ground. Whereas the ground returns for the output stage (and reverb transformer and driver, and LFO) are routed to a bolt at the other end of the chassis, because those circuits are respectively supplied from the screen node cap and reservoir cap.

[jfutral]

That's interesting. This is the first I've heard this. I miss one meeting. Seeing that my preamp filter cap and my preamp ground are on opposite ends of the chassis, would I be better off to A) disconnect the ground where it is, then run a line from the other end of the board to connect with the filter cap ground or B) lift the filter cap ground (which is hog tied with the other filter caps) and run that to connect with the turret board ground at one end or the other? I guess I could try both and see if there isa difference. Or am I too late in the layout and this is best planned in advance of the layout? All I smell right now is solder.

Joe

[jfutral]

How would you feel about tubeswell's suggestion to drop the B+ with zeners to help get the current down? Otherwise I feel like the next step up to try on the resistor is probably 450 ohms, maybe even 500, if I keep trying to drop the dissipation.

BTW, I merely grabbed two 6v6 tubes from my stash o' trash. I've never purchased matched pairs this matched!

Joe

[tubeswell]

Seeing that my preamp filter cap and my preamp ground are on opposite ends of the chassis, would I be better off to A) disconnect the ground where it is, then run a line from the other end of the board to connect with the filter cap ground

The pre-amp ground returns (each by their own separate wires) and the pre-amp filter cap ground return, need to be connected at the same grounding point.

or B) lift the filter cap ground (which is hog tied with the other filter caps) and run that to connect with the turret board ground at one end or the other? I guess I could try both and see if there isa difference. Or am I too late in the layout and this is best planned in advance of the layout? All I smell right now is solder.

If it were me, I would 'unhog-tie' the pre-amp filter cap ground from the other ones (unless you are using a cap can, in which case this won't be possible) and connect that ground return (by its own separate wire) to the other pre-amp ground returns.

The reason for the separate wires all going to one point is that this prevents differences in ground potential (and we are talking micro-differences, but still ones that cause hum havoc nevertheless) interacting between some ground return points but not others - before they connect with the others - thereby exaggerating the differences in ground potential at different points in the ground return path. Remember that each bit of wire is an inductor and a resistor in itself, therefore if you connect some wires together before you connect others, then the ground potential between those parts of the ground return path will be interacting first - potentially creating bigger variations in ground potential (and thus more hum).

[txstrat]

I have found that doing grounding that way results in extremely quiet amps.

tubeswell
Not that my amps are too noisy, actually they are pretty quiet too. But I'm always open to variances. Do you have a layout plan for your Princeton clone or maybe some pics where I can make out the groundings?

[tubeswell]

Sure. Here's a few hi-res shots - so you can zoom in a bit. (Might take a while to download - but quick for me to upload these days ) Excuse the 'messy' look. The lighting in the pics might not be good enough for you to see everything, but I'll try to explain

With the exception of the 10uF filter cap for V1, the separate filter caps are in a doghouse under the chassis, and their ground return wires go separately to various locations in the amp.

The PT has a faraday shield between the windings - this is grounded separately at one of the Pt bolts.

I ran the output stage ground returns and reservoir and screen nodes and the RT ground and 12AT7 grounds and the LFO grounds to the tag terminal strip just at the end of the board. The tag terminals are individually connected to one of the end terminals, which is connected by a short wire to a chassis bolt. The HT CT and reservoir cap ground go to the same tag terminal that is connected to the chassis bolt. The screen node cap ground and grounds for the RT, 12AT7, reverb pan send RCA sleeve, and LFO circuit grounds go to one of the other tag terminals. The bias ground goes to the same one as the bleeder resistors/heater elevation ground. The 6V6 cathodes go to the same one as the 6V6 220k grid load resistors

The preamp ground returns run from the main board (on the back of the board, to the series of eyelets strung together at the input jack end of the board. The input jack ground goes to one of these eyelets. The 10uF filter cap ground goes to the same eyelet as the V1a and v1b ground returns and the vol/TS grounds. The other (22uF) pre-amp filter cap (in the doghouse) goes to the grounds for the PI and PI drivers stages and the reverb recovery stage on the nest eyelet (or two). The 'spare' 22uF filter cap ground also goes to this series of eyelets, but is furthest away from the input jack's ground eyelet. I use 4mm eyelets to fits all these wires in.

All the cable shields (and there were a few of them because of the longer board) go to the pre-amp eyelets as well.

[jfutral]

Well, you guys rock. I love this board. Making me sit here and deconstruct/reconstruct my grounding (physically, not philosophically, though I should probably do that, too) found my hum issue. I metered my grounds and somehow the ground running to my preamp tubes was showing about a .5 ohm resistance. Rewired and reran that ground ala tubeswell's advice and this is easily the quietest amp I've ever made. Not that I've made all that many. This makes my fourth. (let's see. 4 over ten years means 1 every 2.5 years. If I can charge $200-250K apiece, I can retire in about 20 years. Unfortunately at my age, I think that means I'm probably dead by then.) Now I just have some hiss that I know is coming from a POS tube (I've used this tube before and had the same hiss). I just need to buy some new tubes for my friend. Wait. Or have him do it.

I would never have finished any of my amps had it not been for this board.

Thanks everyone!
Joe